2000 character limit reached
Online Updating of Word Representations for Part-of-Speech Tagging
Published 2 Apr 2016 in cs.CL | (1604.00502v1)
Abstract: We propose online unsupervised domain adaptation (DA), which is performed incrementally as data comes in and is applicable when batch DA is not possible. In a part-of-speech (POS) tagging evaluation, we find that online unsupervised DA performs as well as batch DA.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.